English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

In recent years, machine learning algorithm has been widely studied in the field of traffic classification. However, most studies focus on performance improvement of classifier, pro-phase work of traffic classification feature selection is ignored. Therefore, WSU is regarded as metric, an ATFS algorithm (Adaptive threshold feature select) is designed on the basis. Namely, algorithm is based on precision autonomous selection threshold of classifier aiming at different datasets. Each dataset will generate a set of attribute subset eventually. Stable features are selected in different screened attribute subsets through TRF algorithm, thereby reaching the purpose of high precision. The experiment shows that the features finally selected in the algorithm can reach the precision of >96% on C4.5 classifier. Introduction Traffic identification classification is the basis of IP network management and network security monitoring. High-efficient recognition of traffic can assist many ISP (Internet Service Provider) and network administrators to deeply understand traffic composition, thereby controlling the traffic. Traditional traffic identification algorithms are shown as follows: identification based on IP port, identification method based on applied payload signature and identification method based on machine learning. Machine learning is realized based on traffic statistical features. It does not involve payload of message compared with other methods. However, it is still faced with some problems due to data imbalance and concept drift. ATFS and TRF are proposed in the paper to ease data imbalance and concept drift problem. Innovation point in the paper lies in that (1) adaptive linear adjustment threshold is adopted on the basis of literature [4]; (2) an algorithm is proposed for screening attribute subset with stable performance.

A stable feature selection approach for optimizing traffic classification based on adaptive threshold